On the early morning of February 6, OpenAI made a significant announcement regarding an upgrade to the ChatGPT search featureUsers can now directly access the ChatGPT search engine without needing to log into an account.

In the current tide of technology, the field of artificial intelligence (AI) has become a focal point for global competitionLeading technology giants are all vying for an edge in this critical technical revolution that promises to define the contours of our futureHowever, by the end of 2024, a dark horse — DeepSeek — emerged with a force that has sent ripples through the AI industry, sparking deep transformation and widespread discussion.

Revolutionizing Competitiveness

Traditionally, competition within the AI sector focused heavily on model performance and computational power investmentsIn Silicon Valley, major tech players typically employed a “hardware stacking” strategy, continually increasing high-performance GPU capabilities to train more robust AI modelsFor example, OpenAI, together with Nvidia and Microsoft, invested a staggering $500 billion in the "Interstellar Gate" initiative to push forward AI computing infrastructure, securing the leading position for their extensive models.

In contrast, DeepSeek has taken a different approach, centering its advantages on algorithm optimization

Advertisements

Through innovative computational architecture, it has significantly reduced the costs associated with AI trainingDeepSeek has meticulously optimized computation topology, allowing existing computational resources to achieve more efficient data processingAs a result, they have managed to lower training costs by approximately 60% while maintaining model performanceIn simpler terms, while Silicon Valley AI companies might spend $100 to train a model, DeepSeek achieves the same or often better outcomes for only $40. This cost-effective strategy has quickly attracted attention from the global AI market.

Moreover, DeepSeek’s model pricing is impressively competitiveTheir R1 model’s API costs are over 90% lower than OpenAI’s o1-mini, directly drawing in a myriad of users and developer enterprises, rapidly expanding their influence in the global AI arena.

Responses from Industry Giants

Faced with DeepSeek’s aggressive maneuvers, OpenAI quickly reacted to prevent further erosion of its market shareOn February 6, 2025, OpenAI officially unveiled the o3-mini model and announced a reduction in API prices as wellThe o3-mini boasts improvements in accuracy and inference speed compared to its predecessors, while being priced 63% lower than the o1-mini, and a staggering 93% lower than the full version o1. However, even with these adjustments, its pricing remains significantly higher than that of DeepSeek’s R1.

What stands out even more is OpenAI CEO Sam Altman’s admission during a Reddit Q&A session, acknowledging that OpenAI “may have aligned on the wrong side of history” regarding the open-source versus closed-source debate

Advertisements

This statement suggests that OpenAI is reevaluating its closed AI development model, potentially leaning towards a more open approach, which indirectly reflects the profound impact of DeepSeek's open-source strategy on the entire industry, prompting OpenAI to reconsider its stance.

A Paradigm Shift Amidst Crisis

The rise of DeepSeek represents not only a technological breakthrough but also a direct challenge to Silicon Valley's AI business modelIt has illustrated that smarter algorithmic optimization can deliver equivalent, if not superior, AI performance at greatly reduced costs, which poses a significant threat to businesses like OpenAI and Google.

On one hand, it shatters the myth that “computational power reigns supreme.” Previously, the prevailing logic in the AI industry was that “whoever can acquire more high-performance chips will train a stronger model.” DeepSeek’s success emphasizes that optimizing computational architecture significantly decreases AI training costs, enabling smaller enterprises and research institutions to develop powerful AI models, thereby redefining the future trajectory of the AI industry.

On the other hand, the global market could see a major reshufflingThe affordable, high-performance models from DeepSeek are penetrating various AI application scenarios, particularly in the Asian market, where their open-source strategy and pricing advantages have garnered widespread favor among numerous enterprises

Advertisements

Should this trend persist, the dominance of the AI market could undergo significant changesAdditionally, the capital market also feels the tremorsThe competition in the AI industry extends beyond tech, playing out in capital market strategies as well.

DeepSeek's ascent might prompt investors to reassess the value distribution within the AI market, driving capital towards more diverse opportunitiesRecently, Nvidia’s stock saw a rapid decline of 20% within a few days, reflecting growing uncertainty regarding the future landscape of the AI industry.

Efficiency Revolution: The Future Arrives

The rise of DeepSeek signifies a transition in the AI industry from the era of “competing with computational power” to one of “competing with efficiency.” In the past, AI companies battled to see who could invest more money, acquire more GPUs, and train larger modelsNow, it’s about who can train the most powerful models for the least amount of money that will win the marketThis shift not only intensifies competition in the AI sector but also opens the door for more small to medium enterprises to participate in the field.

In the long run, DeepSeek’s achievements indicate a shift in the developmental trajectory of AI technology from mere technical breakthroughs towards engineering optimization and commercial implementationThis transformation will have far-reaching implications for the global AI industry, potentially influencing the formulation of future AI policies and technical standards.

In this silent war of AI, there are no eternal monarchs; only contenders that continue to evolve

Advertisements

Advertisements